Investigation of Control Algorithms For Pv System Under Partial Shading Conditions and Their Effect on The Efficiency of Dc-Dc Converter

Abstract

Generally, conventional energy resources such as fossil fuels are used to meet our electrical energy demand. But the fact is that they are being depleted at a more rapid rate besides creating environmental pollution. In order to mitigate this prob- lem, renewable energy sources such as wind, solar, biomass, hydropower, etc., are used as alternative to produce electricity. Among the renewable energy resources, solar energy has become increasingly popular for many reasons such as low oper- ating cost, no harmful emissions, long operational life, and a clean source. Pho- tovoltaic (PV) panels exhibit non-linear characteristics. Under uniform shading conditions, only one operating point exists where the power is maximum. Under mismatching or partial shading conditions, there exist multiple power peaks. The maximum power point tracking (MPPT) process extracts the maximum available power from the PV panel by fixing the panel voltage corresponding to the maxi- mum power point. A DC-DC converter usually accomplishes this by the impedance matching principle. There are different types of DC-DC converters that are used in between PV panel and the load depending on the applications. This thesis presents a brief literature review on different MPPT methods and losses in the DC-DC converters. The MPPT methods discussed in the literature vary in tracking speed, the number of sensors used, implementation complexity, and their dependence on the PV panels. In addition to different MPPT methods, the literature also presents the loss analysis of the boost converter employed for PV systems. The research gaps were identified based on literature survey, and three objectives have been defined in this thesis. As a first objective, a novel modified current control algorithm is proposed to track the global peak under fast-changing partial shading conditions. The algo- rithm perturbs the operating current in the forward and backward directions. If the irradiance changes during the tracking phase of any irradiance pattern, the proposed algorithm detects the irradiance change and tracks the global peak cor- responding to the new irradiance pattern. The proposed technique uses operating current as a parameter to detect the irradiance change during the tracking process of an existing irradiance pattern. The second objective proposes a global maximum power point tracking (GMPPT) algorithm by perturbing voltage and current. The algorithm perturbs panel volt- age or current based on the value of a control variable. Initially, the operating point is moved to the lowest possible voltage below which there is no global peak. iii Then the perturbation is carried out in the forward direction till the termination criterion is detected. The maximum power is updated during each perturbation so that the global peak is tracked accurately. The proposed technique is compared with two recently published modified GMPPT algorithms with respect to tracking speed and energy efficiency. The third objective compares the losses in the three-level and conventional boost converter for PV applications. Under partial shading conditions, there will be multiple power peaks in power versus voltage (P − V ) characteristics. The location of the global power peak varies over a wide range in the P − V char- acteristics. If the global peak lies to the left side of the P − V curve, the duty cycle required to fix the GMPP would be high. In such cases, the efficiency of the DC-DC converters decreases at higher duty ratios. The study investigates the conduction and switching losses in the conventional boost converter and three- level boost converter through precise mathematical equations. The variation of the losses with respect to the switching frequency is also investigated here. The simulations of all the algorithms and the converters’ loss analysis are performed using MATLAB/Simulink. The two algorithms proposed are compared with recent GMPPT algorithms. The simulations are experimentally validated using an experimental setup. The losses in the three-level boost converter are compared with the conventional boost converter, and the variation of the losses against the switching frequency is plotted and investigated.